PolyFit implements the hypothesis and selection based surface reconstruction method described in
the following paper:
--------------------------------------------------------------
Liangliang Nan and Peter Wonka.
PolyFit: Polygonal Surface Reconstruction from Point Clouds.
ICCV 2017.
--------------------------------------------------------------
Please consider citing the above paper if you use the code/program (or part of it).
====================================================================================================
How to compile the code?
This implementation is based on the following third party libraries (only the version specified
here has been tested):
-) Qt (version 5.8.0). https://www.qt.io/
-) CGAL (version 4.10). http://www.cgal.org/index.html
-) boost (version 1.64. Not a direct dependence, but CGAL relies on it). http://www.boost.org/
Besides, you need either Gurobi or lp_solve for solving the Integer Linear Programs formulated
by our method.
-) Gurobi (version 7.5.1, 64 bit). You may download it and obtain an academic license from
http://www.gurobi.com/
-) lp_solve (version 5.5, source code included in this distribution).
http://lpsolve.sourceforge.net/5.5/
NOTE: Though lp_solve is included and can be used as an alternative solver, it can only solve very
small problems. It is too slow (may take you hours or days) and may fail to solve some
problems. For example, lp_solve takes 15 minutes for the data "Fig1.vg", while Gurobi takes
only 0.02 seconds. So don't doubt to use the Gurobi solver (academic licenses are free) that
is much faster and reliable.
Project file(s) for Visual Studio 2017 are provided.
The code should also be compiled by compilers under other platforms, but not tested yet.
Exectuable/binary files (may not be up-to-date).
Pre-built binary files are available for 64-bit Window users:
http://web.siat.ac.cn/~liangliang/publications/2017/polyfit/polyfit.html
====================================================================================================
How to run the program?
Our algorithm consists of few major steps. To make life easier, this implementation provides a
user interface with a few buttons (with numbered icons in an increasing order) corresponding
to these steps respectively. Please follow the hints on the screen to produce your results.
NOTE: This implementation incorporates a progress logger into the user interface. Thus, running
times should be (slightly) longer than what have been reported in our paper.
====================================================================================================
Data.
Test data can be downloaded from the project page of PolyFit:
http://web.siat.ac.cn/~liangliang/publications/2017/polyfit/polyfit.html
====================================================================================================
License.
This program is free software; you can redistribute it and/or modify it under the terms of the
GNU General Public License as published by the Free Software Foundation; either version 2 of
the License, or (at your option) any later version. The full text of the license can be found
in the accompanying License.txt file.
====================================================================================================
Should you have any questions, comments, or suggestions, please contact me at:
liangliang.nan@gmail.com
Liangliang Nan
http://web.siat.ac.cn/~liangliang/
July 18, 2017
Copyright (C) 2017